First off...I love NetMRG....so here come the problems
1.) ping latency was wacky. First time I set it up it was registering 600m latency when actually is was 0.600msec latency. but apparently that has cleared up now. just odd, and was wondering if anyone else had seen this problem.
2.) I have routers that are seeing Gigs of data thru them according to Netmgr and I know this isn't correct. It appears that NetMrg is just accumulating the data or something, not actually calculating the difference and then graphing it.
3.) forgot one. I set a template graph to a router and when I click on it (when I'm in device tree view) it shows graphs but no data, but with the same template applied to a graph under a specific interface shows data, albeit large amount (Gigs) of data, which is wrong.
netmrg-gather works fine. There are not U's in log file when I run it.
[quotee5e30a2ce4="uNdErDoG"]1.) ping latency was wacky. First time I set it up it was registering 600m latency when actually is was 0.600msec latency. but apparently that has cleared up now. just odd, and was wondering if anyone else had seen this problem.[/quotee5e30a2ce4]
That is odd; our ping script just runs the 'ping' command on the device it's running on. I haven't seen problems with this setup, though; it could be because your box is loaded, or some other incompatibility w/in the ping script.
[quotee5e30a2ce4="uNdErDoG"]2.) I have routers that are seeing Gigs of data thru them according to Netmgr and I know this isn't correct. It appears that NetMrg is just accumulating the data or something, not actually calculating the difference and then graphing it.[/quotee5e30a2ce4]
How did you graph these; using the 'view interface cache' next to your device, or did you apply the monitors yourself?
In either case, for interfaces the monitor types should commonly be 'counter,' as the values continually increase. If they're not set as counters, you'll have to delete the RRDs and start again (after changing the monitor type to 'counter').
[quotee5e30a2ce4="uNdErDoG"]3.) forgot one. I set a template graph to a router and when I click on it (when I'm in device tree view) it shows graphs but no data, but with the same template applied to a graph under a specific interface shows data, albeit large amount (Gigs) of data, which is wrong.[/quotee5e30a2ce4]
Click on 'edit' on the view page (after clicking the orange graph icon). You'll be able to edit which subdevice the template is applied to from here (click 'edit' next to the one that's not working). A template graph will pick the first monitors that match the correct type in the subdevice; so you shouldn't have multiple interfaces being monitored in the same subdevice, for example.
I believe questoins 1 and 2 are now fixed. I did notice after I posted that I had to change the value of the monitor to Absolute or Counter for it to work..thx...
The last problem still is occurring. I have a router and I click on the orange graph icon. I then apply a template to each interface and that all is fine. The name of the interface is there on the graph but no data still. This is across the board on all my routers, I have not tried my switches yet. The template I am applying does not seem to work on any interface though. It's the default network bandwidth utilization snmp template. could an OID be wrong. I did an snmpget and walk and the OID seems fine though. Non of the 64 bit problems I've read earlier.
Got another question. How does a template graph know which SNMP OID I want to use to graph with? You can't set a specific snmp test can you? Maybe I'm missing the point of template. I go into a template graph and it asks to use a "fixed" value or change. If i select change then it only allows me to select from individual graphs I've set up for some interfaces. This seems to defeat the purpose of a template if I have to set a template for every single interface on a router/switch. Am I missing something?
Apparently you're missing something. You only need a template for each type of graph you are creating, not for each interface. Template graphs bind to a sub-device and pull the appropriate monitors to render the graph. When editing a graph, a "fixed" value is used for drawing a static, flat line (such as the bandwidth limit for a type of interface).
that's what I've done. I have a device and then I have several sub-devices. I clicked on the chart next to the device and then clicked "edit" and then added each interface and set the template graph to Network Bandwidth - Utilization, which has two graph items in it's setup. Inbound and Outbound. what I'm asking is how does a template graph "know" what Inbound and Outbound are? How does it know to look at the ifInOctets and ifOutOctets?
Each graph item has several IDs attached to them, the SNMP tests for ifInOctets and ifOutOctets, for example. When you apply a template to a new subdevice, it will create the new monitors using the values from the monitors it is currently set to as graph items. When you view a template that is applied to a subdevice, it looks for the first items in the subdevice that match the correct 'types' of monitors.
Does that help out at all? I wouldn't be surprised if it's still a little unclear.